Opened 14 years ago

Closed 14 years ago

#156 closed defect (fixed)

Two-sided vs One-sided P-values in ROC

Reported by: Peter Owned by: Peter
Priority: major Milestone: yat 0.3 (Public release)
Component: statistics Version: trunk
Keywords: Cc:

Description

I get the same p-value using absolute=true and absolute=false. The doc says these should correspond to tow-tailed and one-tailed test respectively, and if that is true the same p-value should not be returned.

As ROC class has been template for most Score classes, this could be an issue for most Scre classes.

Change History (4)

comment:1 Changed 14 years ago by Jari Häkkinen

Milestone: later0.3 (Public release)

comment:2 Changed 14 years ago by Peter

Status: newassigned

comment:3 Changed 14 years ago by Peter

ROC is redesigned and this no longer an issue see Changeset [821]

Leave this open though because this should be checked in sibling classes as well

comment:4 Changed 14 years ago by Peter

Resolution: fixed
Status: assignedclosed

(In [822]) redesigned tTest fixes #212 and #156

Note: See TracTickets for help on using tickets.